PREVENTING BREAST CANCER: IS IT POSSIBLE?

Ausr. N . Z . J. Surg. 1991,61,884-891

PREVENTING BREAST CANCER: IS IT POSSIBLE?

DAVID INGRAM’ Universiry Department of Surgery, Queen Elizabeth II Medical Centre, Nedlands, Western Australia

Breast cancer is a common cause of suffering and death in Australia. Improvements in detection and therapy will make only small inroads into mortality, and will not reduce suffering. Reducing the incidence of breast cancer by preventive measures is a logical step, and this paper reviews how it might be achieved.

Reducing the population’s exposure to known risk factors for breast cancer is logical; however, the potential for improvement is limited. Reduction in age at fust pregnancy is impractical and measures such as weight reduction and reducing consumption of fat and alcohol are likely to be only partially successful. Increasing activity in youth is a concept that requires further investigation, but is a possible area of promise. Exogenous hormone use probably contributes little to the incidence of breast cancer, but prescribing patterns for postmenopausal oestrogens are changing and require monitoring.

The use of hormone manipulation to prevent breast cancer has considerable potential benefit. Two approaches have been proposed. The f i t involves reversible suppression of ovarian function by luteinizing hormone-releasing hormone (LHRH) agonists after child-bearing has finished, combined with low-dose conjugated equine oestrogen to compensate for oestrogen loss. The other involves the use of the anti-oestrogen, tamoxifen. Both measures have the potential to halve breast cancer incidence. There is some evidence that non-hormonal chemoprevention is possible, but clinical trial data are lacking.

Key words: breast cancer, hormones, prevention.

Introduction Both in importance and in time, health precedes disease, so we ought to consider first how health may be preserved, and then how one may best cure disease.

Galen, AD 200

Approximately one in fifteen Australian women will develop breast cancer by the age of 75, resulting in more than 5000 new cases per year.’ The distress to the patient and her relatives, and the cost to the community of her treatment and for loss of productivity, is enormous. Although most women survive many years after diagnosis, the overall mortality still approaches 40%.

Recent advances have produced some reductions in mortality. Detection of the cancer at an early stage by screening mammography can reduce mortality by as much as 50% for those who have their cancer detected at screening.2 However, less than half of breast cancer cases occur in the 50-70 age

MS, FRCS. FRACS.

Correspondence: D. Ingram, University Department of Sur- gery, Queen Elizabeth I1 Medical Centre, Nedlands, WA 6009, Australia.

Accepted for publication 3 July 1991.

group, for which screening is re~ommended,~ a further 10-40% do not attend when invited to have ~creening ,~ and 20-35% occur as ‘interval’ cancers - that is, they are detected between screen- i n g ~ . ~ Thus, less than 10% of women who develop breast cancer will in fact receive any benefit from screening mammography.

The other major advance is the use of adjuvant systemic therapies, either multi-agent chemotherapy for premenopausal women or tamoxifen for postmenopausal women. Breast cancer patients with involved lymph nodes who receive such therapy have a 25% reduction in mortality.6 However, only about one-third of early breast cancer patients have involved lymph nodes so, again, this modality of treatment will have a relatively small influence on overall mortality.’ The use of adjuvant systemic therapy for node-negative patients remains contro- ~ersial,’.~ but it does appear to give some benefit, for selected groups at least. lo

A more logical approach to the problem is not to reduce the mortality by early diagnosis or by more aggressive therapies as already outlined, but to reduce the incidence of the disease in the population. While inheritance undoubtedly plays a part, it is probable that the great majority of breast cancers have an environmental origin. Such a statement is based on the observation that less than 1O0/o of breast cancer patients have a first-degree relative

PREVENTlNG BREAST CANCER 885

with breast cancer, and on the fact that populations with a low incidence of breast cancer (e.g. the Japanese), after migration to an area with a high incidence (e.g. the United States), have an increasing incidence of breast cancer with increasing dura- tion of settlement, eventually approaching that of the host population. ”

Approaching the problem Several approaches have been proposed with a view to reducing breast cancer incidence:

(1) Identify environmental factors that increase the risk of developing breast cancer - hence, it should be possible to reduce the incidence of breast cancer by reducing the exposure to these risk factors.

(2) Breast cancer development is closely linked to exposure to raised levels of endogenous hormones over a prolonged period of time. By reducing this exposure, it should be possible to reduce considerably the risk of breast cancer developing.

(3) Use non-specific chemopreventive agents, such as antioxidants.

(4) Vaccines.

REDUCING EXPOSURE TO R I S K FACTORS FOR BREAST CANCER DEVELOPMENT

This section reviews recent data that identfy lifestyle- related risk factors which would be open to modi- fication. To these data have been added data from a Table 1. Breast cancer risk factors

Western Australian study we ~ndertook,”-’~ in which risk factors for breast cancer were identified from a community-based population of women, thus identifying what proportion of the population fits into the risk groups. From this, the population- attributable risk - the proportion of all cases attributable to a particular risk factor - was estimated using the formula:

P (R- 1) PR + (1-P)

where R = the relative risk for that factor and P = the population at risk.

The likelihood of achieving change for any particular risk factor is discussed. Risk factor: Age at first pregnancy Women who

delay their first full-term pregnancy have long been recognized as being at increased risk of developing breast cancer. Those women whose first full-term pregnancy is after the age of 25 years (compared with those whose first full-term pregnancy was before 20 years) have a two-fold risk, while for nulliparous women the risk is 2.5 (Table l).”

The mechanism by which age at first pregnancy influences breast cancer development is uncertain; howeter, after the first pregnancy - and, to a less- er extent, after the second pregnancy - the woman’s prolactin concentration falls. Overall, this fall is by a third of the total concentration of prolactin. Women who have a delayed first pregnancy thus have a much longer exposure of their breasts to

Risk factor RR References

Age at fmt pregnancy < 20 cf > 25 years Nulliparity

Weight gain > 10 kg in reproductive years > 60 years and > 80 kg

US non-college team versus college team

Highest quartile Highest quintile

Alcohol consumption > 2 drinkdday > 1 drinWday Any consumption

Oestrogen replacement therapy > 6 years’ oestradiol use

Obesity

Activity in youth

Dietary fat consumption

2. I Lund, 198915 2.5 Lund, 198915

2.1 3.0

Ingram er al., 198912 de Waard, 198216

1.9 Frisch er 01.. 198518

1.6 1.46 Howe et al., 199022

I .3 1.5 Rohan & McMichael, 198fJ2’ 0 Meara er al.. 198929

1.8 Bergkvist er al., 1989-

Ingram et a[:, I 99 I ”

Longnecker er al., 1990”

Oral contraceptive pill Nulliparous, menarche < 13 years, 8-11 years’ use Nulliparous, C 45 years at diagnosis, > 4 years’ use

2.7 Stadel et al., 198834 2.6 McPherson er al., 198733

R R Relative risk

886 INGRAM

high circulating concentrations of prolactin. l4 From the Western Australian control population

data (Table 2), it is apparent that 32% of the population delay their first pregnancy till 25 years or later, and a further I 1 ‘/o are still nulliparous at the age of 30. It can be calculated, then, that the population-attributable risk for age at first pregnancy delayed after the age of 25 is 28% and so, in theory, if all women were to have a pregnancy before the age of 20, breast cancer incidence could be reduced by this amount. With social pressures and financial constraints, however, it is unlikely that any great change in this respect could be achieved.

Table 2. Western Australian control population

Risk factor Population (YO) ___

Age at first pregnancy (years) Nulliparous 18 19-24 25-30 > 30

s60 61-70 71-80

Postmenopausal weight (kg)

> 80

< 10 > 10

< 30

Weight gain in reproductive years (kg)

Proportion of energy from consumed fat (YO)

31-35 36-40 >40

11 7

50 25

I

41 31 12 10

67 33

17 27 30 26

Average alcohol consumption (g/day) 0 42.5 1-9 29 10-29 13 30-49 12.5 3 50 3

Pen- and postmenopausal oestrogen replacement therapy (years of use)

Never 86 1-3 12 4-6 1 7-10 I > 10 0

Parous Oral contraceptive pill use (years)

Never 5.5 1-7 5.0 8-1 I I .o > I I 2.5

Never 36.5 1-7 33.0 8-1 1 10.5 > I 1 6

Nulliparous

Riskfactor: Obesity Obesity appears to play a role in breast cancer development. We have recently demonstrated that women who gain more than 10 kg during their reproductive years have a significant increase in risk for the development of breast cancer, this being in the order of 2.1 .I2 Most, but not all, studies have shown that overweight postmenopausal women have an increased breast cancer risk. This does not seem to apply to premenopausal women.I6 De Waard’s study suggested that, for women over 60 and who weigh more than 80 kg, there is a three-fold increase in risk of the develop ment of breast cancer.

A possible mechanism by which obesity may influence breast cancer development is through the association of obesity with oestrogen binding. With increasing body mass index, there is a progressive fall in sex hormone binding globulin concentration, and a progressive rise in the proportion of oestradiol which is either non-protein-bound or loosely bound to albumin - that is, relatively freely available to the tissues.’* In addition, the lipocytes are an important site of oestrogen production, particularly in postmenopausal women and, as a conse- quence, obese women have higher levels of both oestradiol and oestrone than non-obese women. ”

From our Western Australian data (Table 2), one-third of the population gained more than 10 kg in their reproductive years, resulting in a population- attributable risk of 27%. Ten per cent of postmenopausal women weighed more than 80 kg. As with age at first pregnancy, however, restricting weight gain or losing weight to any significant degree is very difficult to achieve and, realistically, it is unlikely that the incidence of breast cancer could be reduced by influencing this risk factor.

Risk factor: Activity in youth The recent study of US students has identified another interesting risk factor that may have potential for the prevention of breast cancer. It was found that college students who were not in sporting teams had a relative risk of 1.9 for the subsequent development of breast cancer, compared with those students who had been in college teams.”

A possible mechanism is that the increased physical activity of the students in teams would affect breast cancer risk through its effect on reproductive function, as there is evidence that even moderate physical activity at an early age decreases the frequency of ovulatory menstrual cycles;” and in young women, activity depresses luteal progesterone levels.20

From our Western Australian data, we do not have information on the previous level of physical activity but, from the US investigation, approximately half the study group fitted into the category of former athletes. The population-attributable risk is thus calculated at 32%. This area requires further

PREVENTING BREAST CANCER 887

investigation, as there is considerable potential for reducing breast cancer risk by promoting physical activity in young women.

Risk factor: Diet The results of studies to date investigating associations between diet and breast cancer are inconsistent. The original concept that diet, particularly dietary fat, might be related to breast cancer development came from international correlation studies.2’ All have shown significant positive correlations between fat intake and breast cancer mortality. A strong correlation, however, does not mean that the association is causal. Fat intake is, in general, an indicator of affluence, and there are many other differences apparent between countries with high and low breast cancer mortality. National or regional studies have, for the most part, also shown a significant positive correlation between fat intake and breast cancer mortality.

Of individual case-control studies, few have found significant associations between dietary fat and breast cancer risk, although a number of studies identified consumption of fat-containing foods such as meat, butter and margarine, and breast cancer risk. Howe et al., however, recently conducted a combined analysis of the original data from twelve case-control studies.22 Their results showed a consistent, statistically significant, positive association between breast cancer risk and saturated fat consumption in postmenopausal women, the relative risk for highest versus lowest quintiles being 1.46 (P < 0.OOOl). In addition, a consistent protective effect for a number of markers of fruit and veget- able intake was demonstrated. For vitamin C, the relative risk of highest versus lowest quintile was 0.69 (P < 0.OOOl).

At odds with these results are those of the two large North American cohort studies. A study of 89 538 nurses by Willett et al. and of 5485 women by Jones et al. failed to find any evidence of a positive association between breast cancer risk and fat intake.23*24 Indeed, the relative risk for the highest versus lowest quintile of saturated fat in postmenopausal women was only 0.79 in Willett’s study and, similarly, Jones’ study showed an apparent protective effect of high fat intake.

Our own studies have provided some evidence to support the association between dietary fat and breast cancer. In a time-trend study of the UK population, a significant association between breast cancer mortality and fat consumption was indicated, using a 12-year lag i n t e n d z More recently, our case-control study of breast cancer patients indicated associations with fat consumption. There was an increase in risk for total fat consumption (odds ratio [OR] 1.7, confidence limits [c.I.] 0.8- 3.4), saturated fat (OR 1.6, c.1. 0.8-3.1) and mono-unsaturated fat consumption (OR 2.3, c.1. 1.1-4.7), but not for polyunsaturated fats (OR 1.0,

c.1. 0.6-2.1). However, only mono-unsaturated fat consumption reached significance.26 We also showed that increasing saturated fat consumption was significantly correlated with increasing prolactin concentrations, hence suggesting a possible mechanism by which fat consumption may influence breast cancer risk. l4

From our normal population study data, approximately 25% of the female population studied consumed more than 40% of their energy from fat. Calculating the population-attributable risk using our own relative risk figure of 1.7 would give a figure of 15%. Here again, then, there is potential for reducing breast cancer risk by continuing the educational programme promoting a reduction in dietary fat consumption.

Risk factor: Alcohol consumption As with fat consumption, in recent years there has been a pro- fusion of papers investigating alcohol consumption as a risk factor for breast cancer. Most of these have shown that alcohol consumption of more than one drink per day is associated with an increase in the risk of developing breast cancer. A recent meta- analysis based on 26 studies of alcohol and breast cancer suggested the risk was 1.3 (c. 1. 1.1 - 1.5) for two drinks daily.27 Perhaps the data most relevant to us come from the South Australian study, where it was found that consuming more than one alcoholic drink per day was associated with a 1.5-fold increase in risk of breast cancer.28 A more recent British study, however, found no increase in breast cancer risk with alcohol consumption, and suggested that all other studies were of doubtful relevance because of methodological defi~iencies.’~

From our own study data of a control population, more than a quarter of Western Australian women consume, on average, one or more alcoholic drinks per day (based on a drink containing l o g alcohol; Table 2). The population-attributable risk is thus 12%, and so here again there may be potential for reducing breast cancer risk.

Risk factor: Oestrogen replacement therapy A recent study from Scandinavia has provided good documentation of the risks of oestrogen replacement therapy in regard to breast cancer development.30 It was demonstrated that for women with more than six years’ oestradiol use, there was an increase in relative risk for the development of breast cancer of 1.8. Interestingly, there was no apparent increase in risk for those using conjugated equine oestrogen. An editorial in the same issue of the New England Journul of Medicine, while point- ing out some limitations of the study, supported a cautious approach to postmenopausal hormone replacement therapy, and emphasized the need for more data.3’

From our own study data, only 1 YO of postmenopausal women have used oestrogen replacement

888 INGRAM

therapy for more than six years (Table 2). resulting in a population-attributable risk of less than 1 O/O.

This may well change in the future, however, as there is an increasing tendency to prescribe oestrogen replacement therapy to prevent the development of postmenopausal osteoporosis and ischaemic heart disease.

Risk factor: Oral contraceptive pill use While overall there is little evidence that oral contraceptive ill use increases or decreases breast cancer risk.’ in recent years some subgroups have been identified as being at increased risk. In particular, nulliparous women with prolonged oral contraceptive pill use appear to be at increased risk.33*” the order of risk being greater than two-fold (Table 1).

From our data, only 4% of the study population fitted into the group of being nulliparous and of having used the oral contraceptive pill for four or more years (Table 2). The population-attributable risk is thus only in the order of a few per cent, so there is little room for improvement. Furthermore, the formulation of the pill has changed from the time of these studies and it is possible that the risk from the modem contraceptive pill is negligible.

HORMONAL MANIPULATION A N D THE PREVENTION OF BREAST CANCER

Several hormonal intervention-type approaches have been suggested as means of reducing the incidence of breast cancer in our community.

ArtiJicial induction of a pregnancy-like state As discussed previously, an early full-term first pregnancy can considerably reduce the likelihood of a woman developing breast cancer. For obvious reasons, however, it is unrealistic to promote early pregnancy for breast cancer prophylaxis. On the other hand, the concept of trying to deceive the body that a pregnancy has taken place, such as by the use of a short-term high dose of hormonal therapy, warrants further investigation.

Dimethyl benz anthracine (DMBA)-induced breast cancer in rats can be prevented by a full-term pregnancy. In addition, compared with parous rats, virgin rats have an undifferentiated mammary gland with a high proliferative rate in the terminal duct structures. Attempts to differentiate the mammary gland of these animals by exposure to synthetic ovarian hormones, with consequent prevention of breast cancer development, has met with mixed S U C C ~ S S . ~ ~ ~ ~ ~

If breast cancer could be prevented in animals by such a means, extending these studies to humans would be extremely difficult for ethical reasons. The exposure of well young women to high-dose hormone therapy is not without risk.

Artificial induction of menopause Bilateral oophorectomy at an early age considerably reduces breast cancer risk. If done by age 30, the cancer

risk can be reduced by as much as 80%. An approach recently put forward is to induce a reversible inhibition of ovarian function in women once they have finished childbearing, by the use of intermit- tent depot injections of a luteinizing hormone- releasing hormone (LHRH) ag~nis t .~’ This has the additional benefit of providing effective but reversible contraception, and has been estimated to decrease the risk of breast cancer by 50% and of ovarian cancer by 60%. The risk of coronary heart disease and osteoporosis could be obviated, without adding to breast cancer risk, by the use of a low dose of conjugated equine oestrogen for 21 days in each 28- day cycle. Such an approach warrants further con- sideration but, given the current cost of LHRH agonists, would be extremely expensive.

Anti-oestrogens in the prevention of breast cancer Reducing the exposure of the breast epithelium to oestrogens by long-term competitive oestrogen receptor blocking agents such as tamoxifen offers further potential for reducing breast cancer incidence. Recent data from a Swedish trial have shown a 60% reduction in the development of new primary cancer in tamoxifen-treated breast cancer patients.38 Pilot studies are currently under way, investigating the use of tamoxifen in patients at increased risk for developing breast cancer. There is a European Organization for Research on Treatment of Cancer (EORTC) trial of tamoxifen versus close observation for patients with lobular carcinoma in and a study of tamoxifen in women with a first- degree family history of breast cancer.40 The outcome of these pilot studies is awaited with interest.

The use of long-term anti-oestrogen therapy, particularly in otherwise well women, causes some concern. Initially, there was some anxiety that the anti-oestrogen effect might extend to lipid metabolism, resulting in premature coronary heart disease development, and might act on bones, resulting in demineralization and osteoporosis. There is now fairly convincing evidence that these effects do not occur and, if anything, may be prevented by the oestrogen-agonist side affects of tamoxifen. In two studies of long-term tamoxifen use in breast cancer patients, it has become apparent that there is a reduction in cardiac deaths in the tamoxifen-treated group of patient^.^^.^' Animals taking tamoxifen may develop liver tumours, but there is no evidence of this occumng in humans. A real concern relates to endometrial carcinoma. In the Swedish study,38 tamoxifen-treated patients had an increased incidence of endometrial carcinoma, although this effect was not seen in the Scottish study,’” where the women received 20mg daily, compared with 40mg in Sweden. These points must be borne in mind when considering any benefits of therapy.

A new generation of anti-oestrogens that do not have oestrogen-agonist side effects is currently

PREVENTING BREAST CANCER 889

being tested and should reach the market within the next few years. If it can be shown that they do not adversely affect lipid metabolism or bone mineral- ization, and have no other side effects of note, then anti-oestrogen therapy for the high-risk patient may become a reality.

THE USE OF NON-SPECIFIC CHEMOPREVENTIVE AGENTS

Chemoprevention of tumours by the use of agents such as antioxidants has, in theary at least, considerable potential in reducing cancer incidence. Agents such as carotenoids and tocopherols exert their biological activity as quenchers of electroni- cally excited states and as scavengers of free oxy- gen radicals. There is, however, a paucity of data on the role of chemopreventive agents in breast cancer development, and very few clinical trials on their use in breast cancer prevention.

The Finnish Social Insurance Institution’s Mobile Clinic collected blood samples from 23 OOO women in 1968-71. Subsequently, 67 women developed breast cancer and were each matched with two controls. The samples were then analysed for retinol, betacarotene, alpha tocopherol and selenium. The relative risks for retinol and alpha tocopherol were close to 1 .O, but for selenium the risk was 1.7, and for betacarotene, 0.4. The beneficial effect of betacarotene remained statistically significant after adjusting for other variables.42 In addition, the combined analysis of twelve case-control studies by Howe et al. of dietary factors and risk of breast cancer indicated a protective effect for vitamin C and betacarotene consumption, but not retinoLu

Regarding clinical trials, the Instituto Nazionale Tumori in Milan commenced a large chemoprevention trial with the retinoid 4HPR in 1987, with the aim of evaluating any reduction in incidence of contralateral breast cancer in patients already oper- ated on on one side. They aim to accrue more than 2000 patients, and preliminary results should be available in lWl.43

VACCINES

Both RNA and DNA viruses can cause benign and malignant tumours in humans and in animals by a variety of pathogenic mechanisms. Prevention of hepatitis B by vaccination is now feasible, and widespread vaccination of the at-risk population should reduce the incidence of hepatocellular carcinoma. A vaccine against human Epstein-Barr virus carcinoma is under development. The vaccine developed against feline leukaemia virus proves that successful vaccines can be developed against retrovirus infec- tions, which ultimately should include HTLV-1, HTLV-2 and human HIV.

Although there is some evidence that viruses may play a role in breast cancer development, too

little is known to give any hope of prevention by vaccination in the foreseeable future.

Conclusion

REDUCING EXPOSURE TO RISK FACTORS

From a public health point of view, the logical approach is to promote measures that reduce the population’s exposure to risk factors for breast cancer. When each factor is consided in turn, however, it becomes clear that the potential for improvement is somewhat limited.

For a number of sociological, economic and ethical reasons, it is unlikely and even undesirable that age at fmt pregnancy could be moderated. As for weight control, promotional campaigns are unlikely to be more than marginally successful, and would only add to the formidable pressure already exerted by the media upon women to attain the ‘ideal figure’ of an Elle Macpherson! The role of activity in youth has only recently been reported and, although this might certainly prove to be an area of considerable potential for reducing breast cancer risk, con- firmation of the initial data and further research into the mechanisms involved are required before any recommendations can be made.

The role of dietary fat remains controversial: un- certainty still exists as to the importance of the type or amount of fat consumed, and even whether it plays any part at all in breast cancer development. Overall, the evidence of the published data suggests that it does, so it seems reasonable to include breast cancer risk among the reasons for promoting a low- fat diet as being a healthy one. Similarly, although the question of whether alcohol constitutes a risk factor remains debatable, there is sufficient evidence to justify the recommendation that it be consumed in moderation. Promotional programmes such as Western Australia’s ‘drink safe’ campaign should continue.

Finally, regarding exogenous hormone use by either the oral contraceptive pill or menopausal oestrogen replacement therapy, these appear to contrib- ute little to breast cancer incidence. It is sensible, however, to prescribe a low dose of conjugated equine oestrogen for hormone replacement therapy, and not to prolong its use unnecessarily.

Judging from current knowledge, then, it seems that reducing exposure to risk factors for breast cancer will not make a major impact on breast cancer incidence. Nevertheless, the benefits cannot be measured by statistics alone: even small gains are suffering avoided and lives saved.

HORMONE MANIPULATION AND THE PREVENTION OF CANCER

The use of hormone manipulation to prevent breast cancer is much more controversial, but needs to be

890 INGRAM

considered seriously in view of the potential benefits. The artificial induction of a pregnancy-like state

in the late teens could theoretically reduce breast cancer incidence by around one-third, and has the advantage of being, as with vaccination, a once- only procedure. There are, however, considerable research and ethical hurdles to clear before the idea even approaches reality.

The use of LHRH agonists to halt ovarian function after child-bearing has finished, combined with a low dose of conjugated equine oestrogen to compensate the oestrogen loss, would offer considerable benefits beyond the predicted 50% reduction in breast cancer risk. It would also reduce ovarian cancer risk by at least as much. It would provide effective, reversible contraception and obviate the problems of menstruation, thereby reducing the indi- cations for hysterectomy. Moreover, all it would involve is a monthly injection and a nightly ‘pill’. Such an approach clearly warrants testing; unfortu- nately, without a considerable fall in the price of LHRH agonists the cost of trials might be prohibitive.

What of the use of anti-oestrogens to prevent breast cancer? Again, in theory, these have considerable potential, but the paradoxical oestmgen-agonist side effect of tamoxifen makes it undesirable for long-term use. The development of newer agents without this effect may provide long-term therapy for high-risk groups such as women with a family history of breast cancer in first-degree relatives or those with histological changes in the breast such as hyperplasia with atypia or carcinoma in sifu.

The whole population, however, can be considered to be at high risk, since one in fifteen women develops breast cancer at some stage in their lives. We take iodine in our salt to prevent thyroid disease, fluorine in our water to prevent caries, hormones to prevent pregnancy and to prevent heart disease and osteoporosis postmenopausally, to say nothing of numerous other less well-proven sub- stances such as vitamins and calcium; why not an anti-oestrogen to prevent breast cancer? It is the fear of the unknown which, quite rightly, makes us reluctant to accept such an approach. Every child born with a defect, every cancer induced, every patient who develops aplastic anaemia as a result of taking medication, is a disaster. Any move towards widespread use of a new drug, even one which appears to be free of side effects, must be ap- proached with caution.

CHEMOPREVENTION OF BREAST CANCER

Although there are some data to suggest that antioxidants may have a role in preventing breast cancer development, the lack of trial data is a severe limi- tation in deciding what, if any, place such agents have in their widespread use. A major attraction of such agents over the hormonal chemoprevention

measures outlined above is their lack of side effects. Clinical trials are urgently required.

In conclusion, public health measures aimed at reducing exposure to risk factors, although worth- while, are not going to make a major impact on breast cancer incidence. Active intervention would appear to have a greater chance of success. With the hormone approach to intervention, the problems of side effects and cost need to be resolved while, with chemoprevention using antioxidants, paucity of data hinders progress. The outcome of current trials for both means of chemoprevention is awaited. One thing is certain: breast cancer wreaks a terrible toll on our community, and the search must go on for a means to prevent this disease.

I .

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

References McMlcHAEL A. J. & ARMSTRONG B. K. (1988) Breast cancer in Australia: occurrence, risk factors, prevent- ability and screening. Med. J. Ausr. 148, 86-8. FEiG S. A. (1988) Decreased breast cancer mortality through mammographic screening: results of clinical trials. Radiology 167, 659-65. HAITON W. M. & CLARKE-HUNDLEY M. D. (1986) Cancer in Western Australia 1984: An analysis of age and sex specific rates. Statistical Series, no. 2. Health Department of Western Australia, Perth. Western Australia. (1988) First results on mortality reduction in the UK trial of early detection of breast cancer. Lancer ii, 411-16. PEETERS P. H. M., VERBEEK A. L. M., HENDRIKS J. H. C. L., HOLLAND R., MRAVUNAC M. & Voous G. P. (1989) The occurrence of interval cancers in the Nijmegen screening programme. Br. J. Cancer 59,929-32. EARLY BREAST CANCER TRIALISTS’ COLLABORATIVE GROUP (1988) Effects of adjuvant tamoxifen and of cytotoxic therapy on mortality in early breast cancer: an overview of 61 randomised trials among 28,896 women. N. Engl. J. Med. 319, 1681-92. HILL D. J., G u s G. G., RUSSELL I. S., CoLLiNs J. P. & MAPPERSON K. J. (1990) Management of primary operable breast cancer in Victoria. Med. J. Aust.

MCGURE W. L. (1989) Adjuvant therapy of node- negative breast cancer. N. Engl. J. Med. 320,525-7. DEVITA V. T. (1989) Breast cancer therapy: Exercis- ing all our options. N. Engl. J. Med. 320, 527-9. MCGUIRE W. L., TANWN A. K. , ALLRED D. C., CHAMNESS G. C. & CLARK G. M. (1990) How to use prognostic factors in axillary node-negative breast cancer patients. J. Narl Cancer Inst. 82, 1007-15.

epidemiology of cancers of the United States and Japan. Preventative Medicine 6 , 567-94. INGRAM D., NOITAGE E., NG S. , SPARROW L., ROBERTS A. & WiLLcox D. (1989) Obesity and breast disease: the role of the female sex hormones. Cancer 64,

INGRAM D. M., NOITAGE E. M., WILLCOX D. L. &

152.67-72.

WYNDER E. L. & HIRAYAMA T. (1977) Comparative

1049-53.

PREVENTING BREAST CANCER 89 I

ROBERTS A. (1990) Oestrogen binding and risk factors for breast cancer. Br. J. Cancer 61,303-7.

14. INGRAM D. M.. NOITAGE E. M. & ROBERTS A. N. (1990) Prolactin and breast cancer risk. Med. J . Ausr.

15. LUND E. (1989) Reproductive histories and premenopausal breast cancer: different estimates using population, neighbourhood, or hospital controls. Cancer Research 49,4015-19.

16. DE WAARD F. (1982) Nutritional etiology of breast cancer: where are we now, and where are we going? (Editorial). Nurr. Cancer 4. 85-9.

17. GLASS A. R., BURMAN K. D., DAHMS W. T. & BOEHM T. M. (1981) Endocrine function in human obesity. Metabolism 30,89-104.

18. FRISCH R. E., WYSHAK G., ALBRIGHT N. L. er al. (1985) Lower prevalence of breast cancer and cancers of the reproductive system among former college athletes compared to non-athletes. Br. J . Cancer

19. BERNSTEIN L., Ross R. K., LOBO R. A., HANISCH R., KRAILO M. D. & HENDERSON B. E. (1987) The effects of moderate physical activity on menstrual cycle patterns in adolescence: Implications for breast cancer prevention. Br. J. Cancer 55,681-5.

20. ELLISON P. T. & LAGER C. (1986) Moderate recrea- tional running is associated with lowered salivary progesterone profiles in women. Am. J. Obsrer. Cyn- ecol. 154, 1001-3.

21. ARMSTRONG B. & Dou R. (1975) Environmental factors and breast cancer incidence and mortality in different countries, with special reference to dietary practices. Inr. J . Cancer 15,617-31.

Dietary factors and risk of breast cancer: Combined analysis of 12 case-control studies. J. Narl Cancer

23. W I L ~ W. C., STAMPFER M. J., COLDITZ G. A., ROSNER B. A.. HENNEKENS C. H. & S m m F. E. (1987) Dietary fat and the risk of breast cancer. N. Engl. J. Med. 316.22-6.

24. JONES D. Y.. SCHAIZKIN A.. GREEN S . B . er al. (1987) Dietary fat and breast cancer in the National Health & Nutrition Examination Survey 1 epidemiological follow-up study. J. Narl Cancer Insr. 79,465-71.

25. ~ N G R A M D. M. (1981) Trends in diet and breast cancer mortality in England and Wales 1928-1977. Nurr. Cancer 3.75-80.

26. INGRAM D., NOITAGE E. & ROBERTS A. (1991) The role of diet in the development of breast cancer a case-control study of patients with breast cancer, benign epithelial hyperplasia and fibmystic disease of the breast. Br. J . Cancer.

27. LONGNECKER M. P., W i m J. S. & O m M. J. (1990)

153,469-73.

52,885-91.

22. HOWE G . R., HIROHATA T., HISLOP G. er al. (1990)

Imt. 82.561-9.

Meta-analysis of alcohol consumption in relation to risk of breast cancer. 1. Cancer Res. Clin. Oncol.

28. ROHAN T. E. & MCMICHAEL A. J. (1988) Alcohol consumption and risk of breast cancer. Inr. J . Cancer

29. Mwu J., MCPHERSON K., ROBERTS M., JONES L. & VESSEY M. (1989) Alcohol, cigarette smoking and breast cancer. Br. J. Cancer 60.70-3.

S w m C. (1989) The risk of breast cancer after estrogen and estrogen-progestin replacement. N . Engl. J. Med. 321,293-7.

3 1. BARRETT-CONNOR E. .( 1989) Postmenopausal estrogen replacement and breast cancer. N . Engl. J. Med.

32. K ~ E Y J. L. & BERKOWITZG. S. (1988) Breast cancer epidemiology. Cancer Res. 48,5615-23.

33. MCPHERSON K.. VESSEY M. P., NEIL A,, Dou R., JONES L. & ROBERTS M. (1987) Early oral contraceptive use and breast cancer: results of another case- control study. Br. J. Cancer 56.653-60.

34. STADEL B. V., LAI S. H., SCHLESSUMAR J. J. & MURRAY P. (1988) Oral contraceptives and premenopausal breast cancer in nullipamus women. Contra- ception 38,287-99.

35. Russo I. H. & Russo J. (1988) Hormone prevention of mammary carcinogenesis: a new approach in anticancer research. Anticancer Res. 8, 1247-64.

36. Russo I. H. & Russo J. (1986) From pathogenesis to hormone prevention of mammary carcinogenesis. Cancer Surv. 5,649-70.

37. PIKE M. C., Ross R. K., LOBO R. A., KEY T. J. A., P m M. & HENDERSON B. E. (1989) LHRH agonists and the prevention of breast and ovarian cancer. Br. J. Cancer 60, 142-8.

38. FORNANDER T.. CEDERMARK B., MAITSSON A. er al. (1989) Adjuvant Tamoxifen in early breast cancer: occurrence of new primary carcinomas. Lancer i,

39. FENTIMAN I. S. (1989) The endocrine prevention of breast cancer (Editorial). Br. J . Cancer 60, 12-14.

40. Powm T. J., TIUYER C. R., JONES A. L. er al. (1990) Prevention of breast cancer with Tamoxifen: an update on the Royal Marsden Hospital pilot Pro- gramme. Eur. J . Cancer 26,680-4.

41. ROBERTS M. M., ALEXANDER F. E., ANDERSON T. J. er al. (1990) Edinburgh trial of screening for breast cancer: mortality at seven years. Lancer 335,241-6.

42. HAKAMA M., AARAN R. K.. ALPTHAN G. er al. (1990) Blood biochemistry and breast cancer. J. Cancer Res. Clin. Oncol. 116 (Suppl. 2). 1199.

43. COSTA A. (1990) Chemoprevention of breast cancer. J. Cancer Res. Clin. Oncol. 116 (Suppl. 2). 1028.

(Suppl. 2). 854.

41,695-9.

30. BERGKVIST L., ADAM1 H-0.. PERSSON I., HOWER R. &

321. 319-20.

I 17-23.

Documents

PREVENTING BREAST CANCER: IS IT POSSIBLE?